http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
머신러닝을 활용한 EV NVH 최적 모델링에 관한 연구
김용현(Yonghyun Kim),Kanya Benjamin,하형수(Hyeongsoo Ha) 한국자동차공학회 2023 한국자동차공학회 학술대회 및 전시회 Vol.2023 No.11
In the rapidly evolving automotive industry, Noise, Vibration, and Harshness (NVH) control has become a critical aspect of vehicle development. Meeting consumers demands for a quieter and smoother driving experience while adhering to environmental and safety standards requires advanced optimization techniques. This study aims to compare the performance of two NVH optimization algorithms: AVL CAMEO Robust Neural Network and machine learning algorithms from the Python scikit-learn library. To achieve this, a comprehensive evaluation process was conducted. NVH data was collected and preprocessed to suit the requirements of both algorithms. The models were implemented and fine-tuned using appropriate hyperparameters and optimization techniques. Performance evaluation was conducted using test data, focusing on noise, vibration, and stiffness evaluation metrics. The results demonstrated that both AVL CAMEO Robust Neural Network and scikitlearn algorithms offered effective NVH optimization capabilities. However, their strengths and weaknesses differed in various scenarios. AVL CAMEO Robust Neural Network showed promising results in specific NVH aspects, while scikit-learn demonstrated versatility across multiple evaluation metrics. This study provides valuable insights into NVH optimization approaches, contributing to the advancement of vehicle motor development. The findings help automotive manufacturers and researchers choose the most suitable optimization algorithms based on their specific NVH requirements, promoting a smoother and more comfortable driving experience for consumers.